Streamflow and water temperature in the Middle Fork Willamette River (MFWR), western Oregon, have been regulated and altered since the construction of Lookout Point, Dexter, and Hills Creek Dams in 1954 and 1961, respectively. Each year, summer releases from the dams typically are cooler than pre-dam conditions, with the reverse (warmer than pre-dam conditions) occurring in autumn. This pattern has been detrimental to habitat of endangered Upper Willamette River (UWR) Chinook salmon (Oncorhynchus tshawytscha) and UWR winter steelhead (O. mykiss) throughout multiple life stages. In this study, scenarios testing different dam-operation strategies and hypothetical dam-outlet structures were simulated using CE-QUAL-W2 hydrodynamic/temperature models of the MFWR system from Hills Creek Lake (HCR) to Lookout Point (LOP) and Dexter (DEX) Lakes to explore and understand the efficacy of potential flow and temperature mitigation options.

Model scenarios were run in constructed wet, normal, and dry hydrologic calendar years, and designed to minimize the effects of Hills Creek and Lookout Point Dams on river temperature by prioritizing warmer lake surface releases in May–August and cooler, deep releases in September–December. Operational scenarios consisted of a range of modified release rate rules, relaxation of power-generation constraints, variations in the timing of refill and drawdown, and maintenance of different summer maximum lake levels at HCR and LOP. Structural scenarios included various combinations of hypothetical floating outlets near the lake surface and hypothetical new outlets at depth. Scenario results were compared to scenarios using existing operational rules that give temperature management some priority (Base), scenarios using pre-2012 operational rules that prioritized power generation over temperature management (NoBlend), and estimated temperatures from a without-dams condition (WoDams).

Results of the tested model scenarios led to the following conclusions:

• The existing outlets at Lookout Point Dam, because of the range of depths, allow for greater temperature control than the two existing outlets at Hills Creek Dam that are relatively deep.
• Temperature control at HCR through operational scenarios generally was minimal near Hills Creek Dam, but improved downstream toward the head of LOP when decreased release rates held HCR at a low lake elevation year-round.
• Inflows from unregulated streams between HCR and LOP helped to dilute the effects of HCR and achieve more natural stream temperatures before the MFWR entered LOP.
• The relative benefit of any particular scenario depended on the location in the MFWR system used to assess the potential change, with most scenarios involving changes to Hills Creek Dam being less effective with increasing downstream distance, such as downstream of DEX.
• To achieve as much temperature control as the most successful structural scenarios, which were able to resemble without-dam conditions for part of the year, most operational scenarios had to be free of any power-generation requirements at Lookout Point Dam.
• Downstream of DEX, scenarios incorporating a hypothetical floating outlet at either HCR or LOP resulted in similar temperatures, with both scenarios causing a delay in the estimated spring Chinook egg emergence by about 9–10 days compared to base-case temperature-management scenarios.